Projective Holder-Minkowski Colors: A Generalized Set of Commutative & Associative Operations with Inverse Elements for Representing and Manipulating Colors
One of the key problems in dealing with color in rendering, shading, compositing, or image manipulation is that we do not have algebraic structures that support operations over colors. In this paper, we present an all-encompassing framework that can support a set of algebraic structures with associativity, commutativity, and inverse properties. To provide these three properties, we build our algebraic structures on an extension of projective space by allowing for negative and complex numbers. These properties are important for (1) manipulating colors as periodic functions, (2) solving inverse problems dealing with colors, and (3) being consistent with the wave representation of the color. Allowance of negative and complex numbers is not a problem for practical applications, since we can always convert the results into desired range for display purposes as we do in High Dynamic Range imaging. This set of algebraic structures can be considered as a generalization of the Minkowski norm Lp in projective space. These structures also provide a new version of the generalized Holder average with associativity property. Our structures provide inverses of any operation by allowing for negative and complex numbers. These structures provide all properties of the generalized Holder average by providing a continuous bridge between the classical weighted average, harmonic mean, maximum, and minimum operations using a single parameter p.